Dynamic differential annealing-based anti-spoofing model for fingerprint detection using CNN

Data security and privacy play a significant role in human life over the past few years. In the present digital era, advanced technologies utilize wide reliance and ubiquity to assist the counter theft system. Due to the enhanced crime rate, determining the solution becomes a burdensome process to recognize the fingerprint. To overcome such shortcomings, this paper proposes a convolution neural network and dynamic differential annealing (CNN-DDA)-based spoofed fingerprint detection. Here a CNN-DDA approach is proposed to analyze and evaluate the false or forged fingerprint concerning spoof forgery authentication system. The main intention of CNN-DDA architecture employs in investigating a complicated and problematic relationship among various features thus enabling highly detailed features. The proposed CNN-DDA-based spoofed fingerprint detection uses various datasets namely LivDet 2015 and LivDet 2013 for evaluation. Also, the real image set is captured using various fingerprint scanners such as Gelatine, wood glue, ecoflex and modasil. The experimental analysis is conducted for various evaluation measures such as accuracy rate, classification error value rate and processing time. The results revealed that the proposed approach provides high spoofed fingerprint detection with a better accuracy rate, less processing time and classification error.